Progress caring for the chronically ill?: updates from the Annual Research Meeting of Academy Health
Several sessions at the Annual Research Meeting of Academy Health in Chicago were devoted to updating the results of key demonstration projects and pilots related to the care of those with chronic diseases; most notably the Medicare Health Support pilot and the Physician Group Practice Demonstration. The results of these efforts to date highlight the importance of the continued evolution and large scale testing of successful models like HQP’s advanced community-based care management.
Although more data remains to be analyzed, the initial evaluation of the Medicare Health Support pilot presented by RTI International evaluators does not look encouraging either as a way to control costs or improve outcomes. None of the programs participating in the pilot came close to achieving budget neutrality for CMS. Additional analyses with longer follow-up should be coming out soon and merit careful review (especially to see what more can be learned), but this approach does not look promising for Medicare.
The Physician Group Practice Demonstration has been associated with improvements in quality measures and, at least for 4 sites the potential to help control costs. The presenters from RTI International were careful to point out, however, that with regard to cost savings, the research model used in this demonstration is “not a rigorous experimental design”; unlike the Medicare Coordinated Care Demonstration, it is not a randomized trial and there is “incomplete” matching of the comparison group characteristics to those of the intervention group. The presenters noted, “Measuring savings is highly sensitive to target setting methodology, risk adjustment, and demonstration population size.”
Indeed MedPAC also raised this issue in it’s June 2009 Report to Congress: Improving Incentives in the Medicare Program. Beginning on page 258 of the report, referring to the Physician Group Practice Demonstration …
The apparent success of the sites in constraining the rate of cost growth is less clear once risk adjustment effects are taken into account. According to unpublished data from CMS staff, the rate of total expenditure growth without risk adjustment from the base year to performance year 2 were higher or about the same in 8 of the 10 demonstration sites as in their comparison groups. After adjusting for population risk difference (using a methodology similar to the used in Medicare Advantage), only three of the sites had higher total spending growth rates than did their comparison groups. The difference between the unadjusted and adjusted results stems from the fact that 9 of the 10 demonstration sites also reported that their patient risk scores grew faster than risk scores for the sites’ comparison groups. The relatively faster increase in risk scores for the sites may be due to their attracting a greater share of sicker patients than the comparison group, their patients could be getting sicker while enrolled in the demonstration, or the sites may be more fully documenting and coding diagnoses to identify patients for care management and quality improvement initiative. Because the increased risk scores of patients at the sites may be due to improved detection and coding of acute and chronic conditions, actual cost savings in the first two years of the demonstration are unclear.
This is an excellent example of the importance of strengthening our research capacity in this area. Rigorous study designs like that of the Medicare Coordinated Care Demonstration are critical to truly understanding what works to reduce avoidable suffering and cost among our nation’s growing number of older adults with chronic illness. There is no shortcut. The sooner we accept this fact and invest in good research the sooner we will know how to effectively redesign our health care delivery system.
